Samsung logs the biggest chip profit swing in its history as KKR bets $10 billion that dedicated AI infrastructure builders will outcompete hyperscalers.
🗄️ Samsung Q1 2026: Chip Income Jumps 49-Fold to Record $42B Operating Profit
Decoded: Samsung Electronics reported Q1 2026 total revenue of 133.9 trillion won (approximately $97.4 billion) and record operating profit of 57.2 trillion won (approximately $41.6 billion), a 43% increase quarter-on-quarter and an all-time record for the company. The semiconductor division — which supplies DRAM, NAND flash, and high-bandwidth memory — drove the result, posting a 49-fold year-over-year jump in chip income as memory prices surged on AI infrastructure demand. Samsung said it expects server memory demand to remain strong into the second half of 2026 as hyperscalers expand AI adoption and agentic AI deployments accelerate. The company forecast Q2 2026 would be better than Q1 and called 2026 its best year ever. In a notable supply-chain disclosure embedded in the earnings release, Samsung said the global memory supply shortage is projected to worsen through 2027, with advanced HBM supply remaining constrained. Samsung added that in Q2 it plans to reach full utilization of advanced-node production lines and increase HBM4 base-die supply. (Reuters, CNBC, Samsung Global Newsroom, April 30, 2026)
Why it matters: A 49-fold jump in chip income — from a near-zero baseline in the same quarter last year — is the sharpest profitability reversal in Samsung's history, and it arrived when the company is still in a competitive deficit versus SK Hynix on HBM qualification at Nvidia. Samsung's Q1 record despite losing HBM market share signals the AI memory market is large enough to lift even a disadvantaged player to all-time highs. The supply shortage worsening-through-2027 disclosure is the second confirmation of this forecast within two weeks — Nikkei Asia's April 18 report made the same call, with global DRAM production meeting only about 60% of demand. For Micron (MU), Samsung's constrained HBM4 supply schedule and delayed advanced-node ramp extend Micron's window to gain allocation with Nvidia Blackwell and next-generation systems. For Samsung itself, Q2 guidance implying sequential improvement — with HBM4 supply expanding — suggests the company is on track to qualify its memory with Nvidia, just later than originally projected.
💰 KKR Secures $10B to Launch Helix Digital Infrastructure, Hires Ex-AWS CEO Adam Selipsky
Decoded: KKR has secured more than $10 billion to launch a new company called Helix Digital Infrastructure, which will develop and operate artificial intelligence data centers, Bloomberg reported April 30. KKR appointed Adam Selipsky — the former CEO of Amazon Web Services, who led AWS through its period of highest revenue growth before departing in 2023 — as CEO and chair of Helix. Selipsky brings operational experience scaling cloud infrastructure to over $90 billion in annual revenue. Helix will build and operate AI data center campuses, positioning itself between hyperscalers that build infrastructure for internal use and colocation providers that lease space to others. The $10 billion commitment makes Helix one of the largest purpose-built AI infrastructure vehicles ever launched by a private equity firm. The announcement arrived the same week that combined hyperscaler AI capital expenditure — Microsoft, Amazon, Alphabet, and Meta — was disclosed at approximately $700 billion for 2026. (Bloomberg, April 30, 2026)
Why it matters: KKR's $10 billion Helix commitment signals that private capital is moving at scale into the AI infrastructure layer, not just the software layer. Hyperscalers are spending roughly $700 billion combined in 2026 — but that spending is for infrastructure they own and operate for their own cloud customers. Helix is structured to serve enterprises and AI companies that need large-scale compute without building their own hyperscale infrastructure — a segment that CoreWeave and Lambda Labs have targeted, but none has launched with this combination of capital size and operational leadership. Selipsky's AWS tenure is the key credential: he scaled a cloud infrastructure business through the period when AI-driven demand first outpaced supply. The entry of KKR, one of the largest alternative asset managers globally, validates AI data center infrastructure as an investable institutional asset class. Helix will compete for the same power capacity, land, and network interconnects as Microsoft and Amazon.
Stay decoded. See you tomorrow.
— The Get AI Decoded Team
Enjoyed this article?
Subscribe free — AI news decoded for investors, every morning.
No spam. Unsubscribe anytime. Privacy Policy